[Pw_forum] cutoff convergence of sapphire for elastic properties

Nicola Marzari nicola.marzari at epfl.ch
Mon May 7 16:07:52 CEST 2012



Stress tensors tend to be very slow in their convergence - my
understanding is that if one were to compare two approaches -
minimum of the energy as a function of lattice parameter,
or the zero of the trace of the strain, the latter would converge much
more slowly with respect to cutoff.

This slowly convergent contributions were, at least in III-V
semiconductors, fairly independent of the local chemistry, so they
really looked like something rigid (a better description of the part
inside the core).

How much does 1 kbar error translates into an error in lattice
parameter? (keep atoms fixed, using relative coordinates, cutoff fixed,
and expand celldm(1) by 0.3% - what's the change in stress? that change 
should be very well converged)

I'd try to find the equilibrium celldm(1) from the minimum of the 
energy, for a given configuration of the ions (not relaxed, in relative
coordinates), at different cutoffs, to get a sense of what it should be.

Then, at the minimum celldm(1), monitor the stress tensor, to see if it
converges at high cutoff to zero trace.

			nicola



On 07/05/2012 15:56, Jörg Buchwald wrote:
> that was also done.
> JB
>
> Am Mon, 7 May 2012 09:41:35 -0400
> schrieb Iyad AL-QASIR<iyad.ne at gmail.com>:
>
>> Hello,
>> Try first to relax the atomic positions, then relax the structure.
>>
>> Thanks
>>
>>
>> On Mon, May 7, 2012 at 9:22 AM, Jörg Buchwald
>> <joerg.buchwald at iom-leipzig.de
>>> wrote:
>>
>>> Hi,
>>> i tested a relaxed sapphire structure  using a 10x10x10 k point set
>>> with a cutoff energy up to 250 Ry. Even at this Energy, a change of
>>> +/- 5 Ry in the cutoff lead to stresses of the order of 1 kbar.
>>> I used for Al and O pbe-n-van US pseudo potentials from the pseudo
>>> folder of quantum espresso. But also pbe-sp-van for Al and rrkjus
>>> for oxygen led to the same problem.
>>> I don't have a clue how high the cutoff must be, but 250 Ry seems
>>> to be far too high. Is that a problem of the pseudopotentials or
>>> does it have something to do with physical structure of sapphire?
>>> Thx for any hints,
>>> Jörg Buchwald
>>>
>>> --
>>> Leibniz-Institut fuer Oberflaechenmodifizierung e.V.
>>>
>>> Permoserstrasse 15
>>> 04318 Leipzig
>>> GERMANY
>>>
>>> Phone: +49 341 / 235-3367
>>> Web: http://www.uni-leipzig.de/~agmayr
>>>
>>> _______________________________________________
>>> Pw_forum mailing list
>>> Pw_forum at pwscf.org
>>> http://www.democritos.it/mailman/listinfo/pw_forum
>>>
>>
>>
>>
>
>
>


-- 

----------------------------------------------------------------------
Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL


More information about the Pw_forum mailing list